Network Working Group                                        J. Yao, Ed.

Request for Comments: 5336                                   W. Mao, Ed.

Updates: 2821, 2822, 4952                                          CNNIC

Category: Experimental                                    September 2008

          SMTP Extension for Internationalized Email Addresses

Status of This Memo

   This memo defines an Experimental Protocol for the Internet

   community.  It does not specify an Internet standard of any kind.

   Discussion and suggestions for improvement are requested.

   Distribution of this memo is unlimited.

Abstract

   This document specifies an SMTP extension for transport and delivery

   of email messages with internationalized email addresses or header

   information.  Communication with systems that do not implement this

   specification is specified in another document.  This document

   updates some syntaxes and rules defined in RFC 2821 and RFC 2822, and

   has some material updating RFC 4952.

Yao & Mao                     Experimental                      [Page 1]

RFC 5336                   EAI SMTP Extension             September 2008

Table of Contents

   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  3

     1.1.  Role of This Specification . . . . . . . . . . . . . . . .  3

     1.2.  Terminology  . . . . . . . . . . . . . . . . . . . . . . .  3

   2.  Overview of Operation  . . . . . . . . . . . . . . . . . . . .  4

   3.  Mail Transport-Level Protocol  . . . . . . . . . . . . . . . .  4

     3.1.  Framework for the Internationalization Extension . . . . .  4

     3.2.  The UTF8SMTP Extension . . . . . . . . . . . . . . . . . .  5

     3.3.  Extended Mailbox Address Syntax  . . . . . . . . . . . . .  7

     3.4.  The ALT-ADDRESS Parameter  . . . . . . . . . . . . . . . .  9

     3.5.  ALT-ADDRESS Parameter Usage and Response Codes . . . . . . 10

     3.6.  Body Parts and SMTP Extensions . . . . . . . . . . . . . . 11

     3.7.  Additional ESMTP Changes and Clarifications  . . . . . . . 11

       3.7.1.  The Initial SMTP Exchange  . . . . . . . . . . . . . . 12

       3.7.2.  Mail eXchangers  . . . . . . . . . . . . . . . . . . . 12

       3.7.3.  Trace Information  . . . . . . . . . . . . . . . . . . 12

       3.7.4.  UTF-8 Strings in Replies . . . . . . . . . . . . . . . 13

   4.  IANA Considerations  . . . . . . . . . . . . . . . . . . . . . 15

   5.  Security Considerations  . . . . . . . . . . . . . . . . . . . 17

   6.  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 17

   7.  References . . . . . . . . . . . . . . . . . . . . . . . . . . 18

     7.1.  Normative References . . . . . . . . . . . . . . . . . . . 18

     7.2.  Informative References . . . . . . . . . . . . . . . . . . 19

   Appendix A.  Material Updating RFC 4952  . . . . . . . . . . . . . 20

     A.1.  Conventional Message and Internationalized Message . . . . 20

     A.2.  LMTP . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

     A.3.  SMTP Service Extension for DSNs  . . . . . . . . . . . . . 20

     A.4.  Implementation Advice  . . . . . . . . . . . . . . . . . . 20

     A.5.  Applicability of SMTP Extension to Additional Uses . . . . 21

Yao & Mao                     Experimental                      [Page 2]

RFC 5336                   EAI SMTP Extension             September 2008

1.  Introduction

   An internationalized email address includes two parts, the local part

   and the domain part.  The ways email addresses are used by protocols

   are different from the ways domain names are used.  The most critical

   difference is that emails are delivered through a chain of clients

   and servers, while domain names are resolved by name servers looking

   up those names in their own tables.  In addition to this, the Simple

   Mail Transfer Protocol [RFC2821] provides a negotiation mechanism

   about service extension with which clients can discover server

   capabilities and make decisions for further processing.  An extended

   overview of the extension model for internationalized addresses and

   headers appears in [RFC4952], referred to as "the framework document"

   or just as "Framework" elsewhere in this specification.  This

   document specifies an SMTP extension to permit internationalized

   email addresses in envelopes, and UNICODE characters (encoded in

   UTF-8) [RFC3629] in headers.

1.1.  Role of This Specification

   The framework document specifies the requirements for, and describes

   components of, full internationalization of electronic mail.  A

   thorough understanding of the information in that document and in the

   base Internet email specifications [RFC2821] [RFC2822] is necessary

   to understand and implement this specification.

   This document specifies an element of the email internationalization

   work, specifically the definition of an SMTP extension [RFC2821] for

   internationalized email address transport delivery.

1.2.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",

   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this

   document are to be interpreted as described in RFC 2119 [RFC2119].

   The terms "conventional message" and "internationalized message" are

   defined in an appendix to this specification.  The terms "UTF-8

   string" or "UTF-8 character" are used informally to refer to Unicode

   characters encoded in UTF-8 [RFC3629].  All other specialized terms

   used in this specification are defined in the framework document or

   in the base Internet email specifications [RFC2821] [RFC2822].  In

   particular, the terms "ASCII address", "internationalized email

   address", "non-ASCII address", "i18mail address", "UTF8SMTP",

   "message", and "mailing list" are used in this document according to

   the definitions in the framework document.

Yao & Mao                     Experimental                      [Page 3]

RFC 5336                   EAI SMTP Extension             September 2008

   This specification defines only those Augmented BNF (ABNF) [RFC5234]

   syntax rules that are different from those of the base email

   specifications [RFC2821][RFC2822] and, where the earlier rules are

   upgraded or extended, gives them new names.  When the new rule is a

   small modification to the older one, it is typically given a name

   starting with "u".  Rules that are undefined here may be found in the

   base email specifications under the same names.

2.  Overview of Operation

   This specification describes an optional extension to the email

   transport mechanism that permits non-ASCII [ASCII] characters in both

   the envelope and header fields of messages, which are encoded with

   UTF-8 [RFC3629] characters.  The extension is identified with the

   token "UTF8SMTP".  In order to provide information that may be needed

   in downgrading, an optional alternate ASCII address may be needed if

   an SMTP client attempts to transfer an internationalized message and

   encounters a server that does not support this extension.

   The EAI UTF-8 header specification [RFC5335] provides the details of

   how and where non-ASCII characters are permitted in the header fields

   of messages.  The context for this specification is described in the

   framework document.

3.  Mail Transport-Level Protocol

3.1.  Framework for the Internationalization Extension

   The following service extension is defined:

   1.   The name of the SMTP service extension is "Email Address

        Internationalization".

   2.   The EHLO keyword value associated with this extension is

        "UTF8SMTP".

   3.   No parameter values are defined for this EHLO keyword value.  In

        order to permit future (although unanticipated) extensions, the

        EHLO response MUST NOT contain any parameters for that keyword.

        Clients MUST ignore any parameters; that is, clients MUST behave

        as if the parameters do not appear.  If a server includes

        UTF8SMTP in its EHLO response, it MUST be fully compliant with

        this version of this specification.

Yao & Mao                     Experimental                      [Page 4]

RFC 5336                   EAI SMTP Extension             September 2008

   4.   One optional parameter, ALT-ADDRESS, is added to the MAIL and

        RCPT commands of SMTP.  ALT-ADDRESS specifies an all-ASCII

        address which can be used as a substitute for the corresponding

        primary (i18mail) address when downgrading.  More discussion of

        the use of this parameter appears in [RFC4952] and [Downgrade].

   5.   One optional parameter "UTF8REPLY" is added to the VRFY and EXPN

        commands.  The parameter UTF8REPLY has no value.  The parameter

        indicates that the SMTP client can accept Unicode characters in

        UTF-8 encoding in replies from the VRFY and EXPN commands.

   6.   No additional SMTP verbs are defined by this extension.

   7.   Servers offering this extension MUST provide support for, and

        announce, the 8BITMIME extension [RFC1652].

   8.   The reverse-path and forward-path of the SMTP MAIL and RCPT

        commands are extended to allow Unicode characters encoded in

        UTF-8 in mailbox names (addresses).

   9.   The mail message body is extended as specified in [RFC5335].

   10.  The maximum length of MAIL and RCPT command lines is increased

        by 460 characters by the possible addition of the ALT-ADDRESS

        keyword and value.

   11.  The UTF8SMTP extension is valid on the submission port

        [RFC4409].

3.2.  The UTF8SMTP Extension

   An SMTP server that announces this extension MUST be prepared to

   accept a UTF-8 string [RFC3629] in any position in which RFC 2821

   specifies that a mailbox can appear.  That string MUST be parsed only

   as specified in RFC 2821, i.e., by separating the mailbox into source

   route, local part, and domain part, using only the characters colon

   (U+003A), comma (U+002C), and at-sign (U+0040) as specified there.

   Once isolated by this parsing process, the local part MUST be treated

   as opaque unless the SMTP server is the final delivery Mail Transfer

   Agent (MTA).  Any domain names that are to be looked up in the DNS

   MUST first be processed into the form specified in

   "Internationalizing Domain Names in Applications (IDNA)" [RFC3490] by

   means of the ToASCII() operation unless they are already in that

   form.  Any domain names that are to be compared to local strings

   SHOULD be checked for validity and then MUST be compared as specified

   in Section 3.4 of IDNA.

Yao & Mao                     Experimental                      [Page 5]

RFC 5336                   EAI SMTP Extension             September 2008

   An SMTP client that receives the UTF8SMTP extension keyword in

   response to the EHLO command MAY transmit mailbox names within SMTP

   commands as internationalized strings in UTF-8 form.  It MAY send a

   UTF-8 header [RFC5335] (which may also include mailbox names in

   UTF-8).  It MAY transmit the domain parts of mailbox names within

   SMTP commands or the message header as either ACE (ASCII Compatible

   Encoding) labels (as specified in IDNA [RFC3490]) or UTF-8 strings.

   All labels in domain parts of mailbox names which are IDNs (either

   UTF-8 or ACE strings) MUST be valid.  If the original client submits

   a message to a Message Submission Server ("MSA") [RFC4409], it is the

   responsibility of the MSA that all domain labels are valid;

   otherwise, it is the original client's responsibility.  The presence

   of the UTF8SMTP extension does not change the requirement of RFC 2821

   that servers relaying mail MUST NOT attempt to parse, evaluate, or

   transform the local part in any way.

   If the UTF8SMTP SMTP extension is not offered by the Server, the SMTP

   client MUST NOT transmit an internationalized address and MUST NOT

   transmit a mail message containing internationalized mail headers as

   described in [RFC5335] at any level within its MIME structure.  (For

   this paragraph, the internationalized domain name in the form of ACE

   labels as specified in IDNA [RFC3490] is not considered as

   "internationalized".)  Instead, if an SMTP client (SMTP sender)

   attempts to transfer an internationalized message and encounters a

   server that does not support the extension, it MUST make one of the

   following four choices:

   1.  If and only if the SMTP client (sender) is a Message Submission

       Server ("MSA") [RFC4409], it MAY, consistent with the general

       provisions for changes by such servers, rewrite the envelope,

       headers, or message material to make them entirely ASCII and

       consistent with the provisions of RFC 2821 [RFC2821] and RFC 2822

       [RFC2822].

   2.  It may either reject the message during the SMTP transaction or

       accept the message and then generate and transmit a notification

       of non-deliverability.  Such notification MUST be done as

       specified in RFC 2821 [RFC2821], RFC 3464 [RFC3464], and the EAI

       delivery status notification (DSN) specification [RFC5337].

   3.  It may find an alternate route to the destination that permits

       UTF8SMTP.  That route may be discovered by trying alternate Mail

       eXchanger (MX) hosts (using preference rules as specified in RFC

       2821) or using other means available to the SMTP-sender.

   4.  If and only if ASCII addresses are available for all addresses

       that appear in the return path and the specific forward paths

       being attempted, it may downgrade the message to an all-ASCII

Yao & Mao                     Experimental                      [Page 6]

RFC 5336                   EAI SMTP Extension             September 2008

       form as specified in [Downgrade].  An ASCII address is considered

       to be "available" for a particular address if the original

       address in the envelope is in ASCII or if an ALT-ADDRESS

       parameter is specified for a UTF8SMTP address.

   The difference between choice 1 and choice 4 is that choice 1 is

   constrained by Message Submission [RFC4409], while choice 4 is

   constrained by [Downgrade].

3.3.  Extended Mailbox Address Syntax

   RFC 2821, Section 4.1.2, defines the syntax of a mailbox entirely in

   terms of ASCII characters, using the production for a mailbox and

   those productions on which it depends.

   The key changes made by this specification are, informally, to

   o  Change the definition of "sub-domain" to permit either the

      definition above or a UTF-8 string representing a DNS label that

      is conformant with IDNA [RFC3490].

   o  Change the definition of "Atom" to permit either the definition

      above or a UTF-8 string.  That string MUST NOT contain any of the

      ASCII characters (either graphics or controls) that are not

      permitted in "atext"; it is otherwise unrestricted.

   According to the description above, the syntax of an

   internationalized email mailbox name (address) is defined in ABNF

   [RFC5234] as follows.

Yao & Mao                     Experimental                      [Page 7]

RFC 5336                   EAI SMTP Extension             September 2008

           uMailbox = uLocal-part "@" uDomain

             ; Replace Mailbox in RFC 2821, Section 4.1.2

           uLocal-part = uDot-string / uQuoted-string

             ; MAY be case-sensitive

             ; Replace Local-part in RFC 2821, Section 4.1.2

           uDot-string = uAtom *("." uAtom)

             ; Replace Dot-string in RFC 2821, Section 4.1.2

           uAtom = 1*ucharacter

                 ; Replace Atom in RFC 2821, Section 4.1.2

           ucharacter = atext / UTF8-non-ascii

           atext = <See Section 3.2.4 of RFC 2822>

           uQuoted-string = DQUOTE *uqcontent DQUOTE

             ; Replace Quoted-string in RFC 2821, Section 4.1.2

           DQUOTE = <See appendix B.1 of RFC 5234>

           uqcontent = qcontent / UTF8-non-ascii

           qcontent = <See Section 3.2.5 of RFC 2822>

           uDomain = (sub-udomain 1*("." sub-udomain)) / address-literal

             ; Replace Domain in RFC 2821, Section 4.1.2

           address-literal = <See Section 4.1.2 of RFC 2822>

           sub-udomain = uLet-dig [uLdh-str]

             ; Replace sub-domain in RFC 2821, Section 4.1.2

           uLet-dig = Let-dig / UTF8-non-ascii

           Let-dig = <See Section 4.1.3 of RFC 2821>

           uLdh-str = *( ALPHA / DIGIT / "-" / UTF8-non-ascii) uLet-dig

             ; Replace Ldh-str in RFC 2821, Section 4.1.3

           UTF8-non-ascii = UTF8-2 / UTF8-3 / UTF8-4

           UTF8-2 =  <See Section 4 of RFC 3629>

           UTF8-3 =  <See Section 4 of RFC 3629>

           UTF8-4 =  <See Section 4 of RFC 3629>

Yao & Mao                     Experimental                      [Page 8]

RFC 5336                   EAI SMTP Extension             September 2008

   The value of "uDomain" SHOULD be verified by applying the tests

   specified as part of IDNA [RFC3490].  If that verification fails, the

   email address with that uDomain MUST NOT be regarded as a valid email

   address.

3.4.  The ALT-ADDRESS Parameter

   If the UTF8SMTP extension is offered, the syntax of the SMTP MAIL and

   RCPT commands is extended to support the optional esmtp-keyword "ALT-

   ADDRESS".  That keyword specifies an alternate all-ASCII address that

   may be used when downgrading.  If the ALT-ADDRESS esmtp-keyword is

   used, it MUST have an associated esmtp-value (ALT-ADDRESS-esmtp-

   value, which is defined below).

   While it may be tempting to consider ALT-ADDRESS as a general-purpose

   second-chance address, such behavior is not defined here.  Instead,

   in this specification ALT-ADDRESS only has meaning when the

   associated primary address is non-ASCII and the message is

   downgraded.  This restriction allows for future extension of the

   specification even though no such extensions are currently

   anticipated.

   Based on the definition of mail-parameters in [RFC2821], the ALT-

   ADDRESS parameter usage in the commands of MAIL and RCPT is defined

   as follows.  The following definitions are given in the same format

   as used in RFC 2821.

        "MAIL FROM:" ("<>" / uReverse-path) [ SP Mail-parameters ] CRLF

           ; Update the MAIL command in RFC 2821, Section 4.1.1.2.

           ; A new parameter defined by the ABNF non-terminal

           ; <ALT-ADDRESS-parameter> is added.  It complies

           ; with the syntax specified for <esmtp-param> in RFC 2821.

        "RCPT TO:" ("<Postmaster@" uDomain ">" / "<Postmaster>" /

              uForward-path) [ SP Rcpt-parameters ] CRLF

               ; Update RCPT command in RFC 2821, Section 4.1.1.3.

               ; A new parameter defined by the ABNF non-terminal

               ; <ALT-ADDRESS-parameter> is added.  It complies

               ; with the syntax specified for <esmtp-param>.

               ; uDomain is defined in Section 3.3 of this document.

        uReverse-path = uPath

           ; Replace Reverse-path in RFC 2821, Section 4.1.2.

        uForward-path = uPath

           ; Replace Forward-path in RFC 2821, Section 4.1.2.

Yao & Mao                     Experimental                      [Page 9]

RFC 5336                   EAI SMTP Extension             September 2008

        uPath = "<" [ A-d-l ":" ] uMailbox ">"

           ; Replace Path in RFC 2821, Section 4.1.2.

           ; uMailbox is defined in Section 3.3 of this document.

        A-d-l = <See Section 4.1.2 of RFC 2821>

        ALT-ADDRESS-parameter = "ALT-ADDRESS=" ALT-ADDRESS-value

        ALT-ADDRESS-value = xtext

               ; The value is a mailbox name encoded as xtext.

        xtext = <See Section 4.2 of RFC 3461>

   The ALT-ADDRESS-parameter MUST NOT appear more than once in any MAIL

   or RCPT command.  ALT-ADDRESS-esmtp-value MUST be an all-ASCII email

   address before xtext encoding.

3.5.  ALT-ADDRESS Parameter Usage and Response Codes

   An "internationalized message" as defined in the appendix of this

   specification MUST NOT be sent to an SMTP server that does not

   support UTF8SMTP.  Such a message MAY be rejected by a server if it

   lacks ALT-ADDRESSes as discussed in Section 3.2 of this

   specification.

   The three-digit reply codes used in this section are consistent with

   their meanings as defined in RFC 2821.

   When messages are rejected because the RCPT command requires an ALT-

   ADDRESS, the response code 553 is used with the meaning "mailbox name

   not allowed".  When messages are rejected for other reasons, such as

   the MAIL command requiring an ALT-ADDRESS, the response code 550 is

   used with the meaning "mailbox unavailable".  When the server

   supports enhanced mail system status codes [RFC3463], response code

   "X.6.7" [RFC5248] is used, meaning that "The ALT-ADDRESS is required

   but not specified".

   If the response code is issued after the final "." of the DATA

   command, the response code "554" is used with the meaning

   "Transaction failed".  When the server supports enhanced mail system

   status codes [RFC3463], response code "X.6.9" [RFC5248] is used,

   meaning that "UTF8SMTP downgrade failed".

Yao & Mao                     Experimental                     [Page 10]

RFC 5336                   EAI SMTP Extension             September 2008

3.6.  Body Parts and SMTP Extensions

   There is no ESMTP parameter to assert that a message is an

   internationalized message.  An SMTP server that requires accurate

   knowledge of whether a message is internationalized is required to

   parse all message header fields and MIME header fields in the message

   body.

   While this specification requires that servers support the 8BITMIME

   extension [RFC1652] to ensure that servers have adequate handling

   capability for 8-bit data and to avoid a number of complex encoding

   problems, the use of internationalized addresses obviously does not

   require non-ASCII body parts in the MIME message.  The UTF8SMTP

   extension MAY be used with the BODY=8BITMIME parameter if that is

   appropriate given the body content or, with the BODY=BINARYMIME

   parameter, if the server advertises BINARYMIME [RFC3030] and that is

   appropriate.

   Assuming that the server advertises UTF8SMTP and 8BITMIME, and

   receives at least one non-ASCII address, with or without ALT-ADDRESS,

   the precise interpretation of 'No BODY parameter', "BODY=8BITMIME",

   and "BODY=BINARYMIME" in the MAIL command is:

   1.  If there is no BODY parameter, the header contains UTF-8

       characters, but all the body parts are in ASCII (possibly as the

       result of a content-transfer-encoding).

   2.  If a BODY=8BITMIME parameter is present, the header contains

       UTF-8 characters, and some or all of the body parts contain 8-bit

       line-oriented data.

   3.  If a BODY=BINARYMIME parameter is present, the header contains

       UTF-8 characters, and some or all body parts contain binary data

       without restriction as to line lengths or delimiters.

3.7.  Additional ESMTP Changes and Clarifications

   The information carried in the mail transport process involves

   addresses ("mailboxes") and domain names in various contexts in

   addition to the MAIL and RCPT commands and extended alternatives to

   them.  In general, the rule is that, when RFC 2821 specifies a

   mailbox, this specification expects UTF-8 to be used for the entire

   string; when RFC 2821 specifies a domain name, the name SHOULD be in

   the form of ACE labels if its raw form is non-ASCII.

   The following subsections list and discuss all of the relevant cases.

Yao & Mao                     Experimental                     [Page 11]

RFC 5336                   EAI SMTP Extension             September 2008

3.7.1.  The Initial SMTP Exchange

   When an SMTP connection is opened, the server normally sends a

   "greeting" response consisting of the 220 response code and some

   information.  The client then sends the EHLO command.  Since the

   client cannot know whether the server supports UTF8SMTP until after

   it receives the response from EHLO, any domain names that appear in

   this dialogue, or in responses to EHLO, MUST be in the hostname form,

   i.e., internationalized ones MUST be in the form of ACE labels.

3.7.2.  Mail eXchangers

   Organizations often authorize multiple servers to accept mail

   addressed to them.  For example, the organization may itself operate

   more than one server, and may also or instead have an agreement with

   other organizations to accept mail as a backup.  Authorized servers

   are generally listed in MX records as described in RFC 2821.  When

   more than one server accepts mail for the domain-part of a mailbox,

   it is strongly advised that either all or none of them support the

   UTF8SMTP extension.  Otherwise, surprising downgrades can happen

   during temporary failures, which users might perceive as a serious

   reliability issue.

3.7.3.  Trace Information

   When an SMTP server receives a message for delivery or further

   processing, it MUST insert trace ("time stamp" or "Received")

   information at the beginning of the message content.  "Time stamp" or

   "Received" appears in the form of "Received:" lines.  The most

   important use of Received: lines is for debugging mail faults.  When

   the delivery SMTP server makes the "final delivery" of a message, it

   inserts a Return-path line at the beginning of the mail data.  The

   primary purpose of the Return-path is to designate the address to

   which messages indicating non-delivery or other mail system failures

   are to be sent.  For the trace information, this memo updates the

   time stamp line and the return path line [RFC2821] formally defined

   as follows:

      uReturn-path-line = "Return-Path:" FWS uReverse-path <CRLF>

          ; Replaces Return-path-line in Section 4.4 of RFC 2821

          ; uReverse-path is defined in Section 3.3 of this document

      uTime-stamp-line = "Received:" FWS uStamp <CRLF>

          ; Replaces Time-stamp-line in Section 4.4 of RFC 2821

      uStamp = From-domain By-domain uOpt-info ";"  FWS date-time

          ; Replaces Stamp in Section 4.4 of RFC 2821

Yao & Mao                     Experimental                     [Page 12]

RFC 5336                   EAI SMTP Extension             September 2008

       uOpt-info = [Via] [With] [ID] [uFor]

          ; Replaces Opt-info in Section 4.4 of RFC 2821

          ; The protocol value for With will allow a UTF8SMTP value

         uFor = "FOR" ( FWS (uPath / uMailbox) ) CFWS

          ; Replaces For in Section 4.4 of RFC 2821

          ; uPath and uMailbox are defined in Sections 2.4 and

          ; 2.3, respectively, of this document

   Note: The FOR parameter has been changed to match the definition in

   [RFC2821bis], permitting only one address in the For clause.  The

   group working on that document reached mailing list consensus that

   the syntax in [RFC2821] that permitted more than one address was

   simply a mistake.

   Except in the 'uFor' clause and 'uReverse-path' value where non-ASCII

   domain names may be used, internationalized domain names in Received

   fields MUST be transmitted in the form of ACE labels.  The protocol

   value of the WITH clause when this extension is used is one of the

   UTF8SMTP values specified in the "IANA Considerations" section of

   this document.

3.7.4.  UTF-8 Strings in Replies

3.7.4.1.  MAIL and RCPT Commands

   If the client issues a RCPT command containing non-ASCII characters,

   the SMTP server is permitted to use UTF-8 characters in the email

   address associated with 251 and 551 response codes.

   If an SMTP client follows this specification and sends any RCPT

   commands containing non-ASCII addresses, it MUST be able to accept

   and process 251 or 551 responses containing UTF-8 email addresses.

   If a given RCPT command does not include a non-ASCII envelope

   address, the server MUST NOT return a 251 or 551 response containing

   a non-ASCII mailbox.  Instead, it MUST transform such responses into

   250 or 550 responses that do not contain addresses.

3.7.4.2.  VRFY and EXPN Commands and the UTF8REPLY Parameter

   If the VRFY and EXPN commands are transmitted with an optional

   parameter "UTF8REPLY", it indicates the client can accept UTF-8

   strings in replies from those commands.  This allows the server to

   use UTF-8 strings in mailbox names and full names that occur in

   replies without concern that the client might be confused by them.

   An SMTP client that conforms to this specification MUST accept and

   correctly process replies from the VRFY and EXPN commands that

   contain UTF-8 strings.  However, the SMTP server MUST NOT use UTF-8

Yao & Mao                     Experimental                     [Page 13]

RFC 5336                   EAI SMTP Extension             September 2008

   strings in replies if the SMTP client does not specifically allow

   such replies by transmitting this parameter.  Most replies do not

   require that a mailbox name be included in the returned text, and

   therefore UTF-8 is not needed in them.  Some replies, notably those

   resulting from successful execution of the VRFY and EXPN commands, do

   include the mailbox, making the provisions of this section important.

   VERIFY (VRFY) and EXPAND (EXPN) command syntaxes are changed to:

       "VRFY" SP (uLocal-part / uMailbox) [SP "UTF8REPLY"] CRLF

              ; uLocal-part and uMailbox are defined in

              ; Section 3.3 of this document.

       "EXPN" SP ( uLocal-part / uMailbox ) [ SP "UTF8REPLY" ] CRLF

              ; uLocal-part and uMailbox are defined in

              ; Section 3.3 of this document.

   The "UTF8REPLY" parameter does not use a value.  If the reply to a

   VERIFY (VRFY) or EXPAND (EXPN) command requires UTF-8, but the SMTP

   client does not use the "UTF8REPLY" parameter, then the server MUST

   use either the response code 252 or 550.  Response code 252, defined

   in [RFC2821], means "Cannot VRFY user, but will accept the message

   and attempt the delivery".  Response code 550, also defined in

   [RFC2821], means "Requested action not taken: mailbox unavailable".

   When the server supports enhanced mail system status codes [RFC3463],

   the enhanced response code as specified below is used.  Using the

   "UTF8REPLY" parameter with a VERIFY (VRFY) or EXPAND (EXPN) command

   enables UTF-8 replies for that command only.

   If a normal success response (i.e., 250) is returned, the response

   MAY include the full name of the user and MUST include the mailbox of

   the user.  It MUST be in either of the following forms:

         User Name <uMailbox>

            ; uMailbox is defined in Section 3.3 of this document.

            ; User Name can contain non-ASCII characters.

         uMailbox

            ; uMailbox is defined in Section 3.3 of this document.

   If the SMTP reply requires UTF-8 strings, but UTF-8 is not allowed in

   the reply, and the server supports enhanced mail system status codes

   [RFC3463], the enhanced response code is either "X.6.8" or "X.6.10"

   [RFC5248], meaning "A reply containing a UTF-8 string is required to

   show the mailbox name, but that form of response is not permitted by

   the client".

Yao & Mao                     Experimental                     [Page 14]

RFC 5336                   EAI SMTP Extension             September 2008

   If the SMTP client does not support the UTF8SMTP extension, but

   receives a UTF-8 string in a reply, it may not be able to properly

   report the reply to the user, and some clients might crash.

   Internationalized messages in replies are only allowed in the

   commands under the situations described above.  Under any other

   circumstances, UTF-8 text MUST NOT appear in the reply.

   Although UTF-8 is needed to represent email addresses in responses

   under the rules specified in this section, this extension does not

   permit the use of UTF-8 for any other purposes.  SMTP servers MUST

   NOT include non-ASCII characters in replies except in the limited

   cases specifically permitted in this section.

4.  IANA Considerations

   IANA has added a new value "UTF8SMTP" to the SMTP Service Extension

   subregistry of the Mail Parameters registry, according to the

   following data:

        +----------+---------------------------------+-----------+

        | Keywords | Description                     | Reference |

        +----------+---------------------------------+-----------+

        | UTF8SMTP | Internationalized email address | [RFC5336] |

        +----------+---------------------------------+-----------+

   This document adds new values to the SMTP Enhanced Status Code

   subregistry of the Mail Parameters registry, following the guidance

   in Sections 3.5 and 3.7.4.2 of this document, and being based on

   [RFC5248].  The registration data is as follows:

Yao & Mao                     Experimental                     [Page 15]

RFC 5336                   EAI SMTP Extension             September 2008

      Code:               X.6.7

      Sample Text:        The ALT-ADDRESS is required but not specified

      Associated basic status code:  553, 550

      Description:        This indicates the reception of a MAIL or RCPT

                          command that required an ALT-ADDRESS parameter

                          but such parameter was not present.

      Defined:            RFC 5336  (Experimental track)

      Submitter:          Jiankang YAO

      Change controller:  IESG.

      Code:               X.6.8

      Sample Text:        UTF-8 string reply is required,

                          but not permitted by the client

      Associated basic status code:  553, 550

      Description:        This indicates that a reply containing a UTF-8

                          string is required to show the mailbox name,

                          but that form of response is not

                          permitted by the client.

      Defined:            RFC  5336.  (Experimental track)

      Submitter:          Jiankang YAO

      Change controller:  IESG.

       Code:               X.6.9

       Sample Text:        UTF8SMTP downgrade failed

       Associated basic status code:  550

       Description:        This indicates that transaction failed

                           after the final "." of the DATA command.

       Defined:            RFC  5336.  (Experimental track)

       Submitter:          Jiankang YAO

       Change controller:  IESG.

      Code:               X.6.10

      Sample Text:        UTF-8 string reply is required,

                          but not permitted by the client

      Associated basic status code:  252

      Description:        This indicates that a reply containing a UTF-8

                          string is required to show the mailbox name,

                          but that form of response is not

                          permitted by the client.

      Defined:            RFC 5336.  (Experimental track)

      Submitter:          Jiankang YAO

      Change controller:  IESG.

Yao & Mao                     Experimental                     [Page 16]

RFC 5336                   EAI SMTP Extension             September 2008

   The "Mail Transmission Types" registry under the Mail Parameters

   registry is requested to be updated to include the following new

   entries:

   +---------------+----------------------------+----------------------+

   | WITH protocol | Description                | Reference            |

   | types         |                            |                      |

   +---------------+----------------------------+----------------------+

   | UTF8SMTP      | UTF8SMTP with Service      | [RFC5336]            |

   |               | Extensions                 |                      |

   | UTF8SMTPA     | UTF8SMTP with SMTP AUTH    | [RFC4954] [RFC5336]  |

   | UTF8SMTPS     | UTF8SMTP with STARTTLS     | [RFC3207] [RFC5336]  |

   | UTF8SMTPSA    | UTF8SMTP with both         | [RFC3207] [RFC4954]  |

   |               | STARTTLS and SMTP AUTH     | [RFC5336]            |

   +---------------+----------------------------+----------------------+

5.  Security Considerations

   See the extended security considerations discussion in the framework

   document [RFC4952].

6.  Acknowledgements

   Much of the text in the initial version of this specification was

   derived or copied from [Emailaddr] with the permission of the author.

   Significant comments and suggestions were received from Xiaodong LEE,

   Nai-Wen Hsu, Yangwoo KO, Yoshiro YONEYA, and other members of the JET

   team and were incorporated into the specification.  Additional

   important comments and suggestions, and often specific text, were

   contributed by many members of the WG and design team.  Those

   contributions include material from John C Klensin, Charles Lindsey,

   Dave Crocker, Harald Tveit Alvestrand, Marcos Sanz, Chris Newman,

   Martin Duerst, Edmon Chung, Tony Finch, Kari Hurtta, Randall Gellens,

   Frank Ellermann, Alexey Melnikov, Pete Resnick, S. Moonesamy, Soobok

   Lee, Shawn Steele, Alfred Hoenes, Miguel Garcia, Magnus Westerlund,

   and Lars Eggert.  Of course, none of the individuals are necessarily

   responsible for the combination of ideas represented here.

Yao & Mao                     Experimental                     [Page 17]

RFC 5336                   EAI SMTP Extension             September 2008

7.  References

7.1.  Normative References

   [ASCII]       American National Standards Institute (formerly United

                 States of America Standards Institute), "USA Code for

                 Information Interchange", ANSI X3.4-1968, 1968.

   [RFC1652]     Klensin, J., Freed, N., Rose, M., Stefferud, E., and D.

                 Crocker, "SMTP Service Extension for 8bit-

                 MIMEtransport", RFC 1652, July 1994.

   [RFC2119]     Bradner, S., "Key words for use in RFCs to Indicate

                 Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC2821]     Klensin, J., "Simple Mail Transfer Protocol", RFC 2821,

                 April 2001.

   [RFC2822]     Resnick, P., "Internet Message Format", RFC 2822,

                 April 2001.

   [RFC3461]     Moore, K., "Simple Mail Transfer Protocol (SMTP)

                 Service Extension for Delivery Status Notifications

                 (DSNs)", RFC 3461, January 2003.

   [RFC3463]     Vaudreuil, G., "Enhanced Mail System Status Codes",

                 RFC 3463, January 2003.

   [RFC3464]     Moore, K. and G. Vaudreuil, "An Extensible Message

                 Format for Delivery Status Notifications", RFC 3464,

                 January 2003.

   [RFC3490]     Faltstrom, P., Hoffman, P., and A. Costello,

                 "Internationalizing Domain Names in Applications

                 (IDNA)", RFC 3490, March 2003.

   [RFC3629]     Yergeau, F., "UTF-8, a transformation format of ISO

                 10646", STD 63, RFC 3629, November 2003.

   [RFC4409]     Gellens, R. and J. Klensin, "Message Submission for

                 Mail", RFC 4409, April 2006.

   [RFC4952]     Klensin, J. and Y. Ko, "Overview and Framework for

                 Internationalized Email", RFC 4952, July 2007.

   [RFC5234]     Crocker, D. and P. Overell, "Augmented BNF for Syntax

                 Specifications: ABNF", STD 68, RFC 5234, January 2008.

Yao & Mao                     Experimental                     [Page 18]

RFC 5336                   EAI SMTP Extension             September 2008

   [RFC5248]     Hansen, T. and J. Klensin, "A Registry for SMTP

                 Enhanced Mail System Status Codes", BCP 138, RFC 5248,

                 June 2008.

   [RFC5335]     Abel, Y., Ed., "Internationalized Email Headers",

                 RFC 5335, September 2008.

   [RFC5337]     Newman, C. and A. Melnikov, Ed., "Internationalized

                 Delivery Status and Disposition Notifications",

                 RFC 5337, September 2008.

7.2.  Informative References

   [Downgrade]   Fujiwara, K. and Y. Yoneya, "Downgrading mechanism for

                 Email Address Internationalization", Work in Progress,

                 July 2008.

   [Emailaddr]   Klensin, J., "Internationalization of Email Addresses",

                 Work in Progress, July 2005.

   [RFC0974]     Partridge, C., "Mail routing and the domain system",

                 RFC 974, January 1986.

   [RFC2033]     Myers, J., "Local Mail Transfer Protocol", RFC 2033,

                 October 1996.

   [RFC2821bis]  Klensin, J., "Simple Mail Transfer Protocol", Work

                 in Progress, July 2008.

   [RFC3030]     Vaudreuil, G., "SMTP Service Extensions for

                 Transmission of Large and Binary MIME Messages",

                 RFC 3030, December 2000.

   [RFC3207]     Hoffman, P., "SMTP Service Extension for Secure SMTP

                 over Transport Layer Security", RFC 3207,

                 February 2002.

   [RFC4954]     Siemborski, R., Ed. and A. Melnikov, Ed., "SMTP Service

                 Extension for Authentication", RFC 4954, July 2007.

Yao & Mao                     Experimental                     [Page 19]

RFC 5336                   EAI SMTP Extension             September 2008

Appendix A.  Material Updating RFC 4952

   RFC 4952, the overview and framework document covering this set of

   extensions for internationalized email, was completed before this

   specification, which specifies a particular part of the protocol set.

   This appendix, which is normative, contains material that would have

   been incorporated into RFC 4952 had it been delayed until the work

   described in the rest of this specification was completed.  This

   material should be included in any update to RFC 4952.

A.1.  Conventional Message and Internationalized Message

   o  A conventional message is one that does not use any extension

      defined in this document or in the UTF-8 header specification

      [RFC5335], and which is strictly conformant to RFC 2822 [RFC2822].

   o  An internationalized message is a message utilizing one or more of

      the extensions defined in this specification or in the UTF-8

      header specification [RFC5335], so that it is no longer conformant

      to the RFC 2822 specification of a message.

A.2.  LMTP

   LMTP [RFC2033] may be used as the final delivery agent.  In such

   cases, LMTP may be arranged to deliver the mail to the mail store.

   The mail store may not have UTF8SMTP capability.  LMTP needs to be

   updated to deal with these situations.

A.3.  SMTP Service Extension for DSNs

   The existing Draft Standard regarding delivery status notifications

   (DSNs) [RFC3461] is limited to ASCII text in the machine readable

   portions of the protocol.  "International Delivery Status and

   Disposition Notifications" [RFC5337] adds a new address type for

   international email addresses so an original recipient address with

   non-ASCII characters can be correctly preserved even after

   downgrading.  If an SMTP server advertises both the UTF8SMTP and the

   DSN extension, that server MUST implement EAI DSN [RFC5337] including

   support for the ORCPT parameter.

A.4.  Implementation Advice

   In the absence of this extension, SMTP clients and servers are

   constrained to using only those addresses permitted by RFC 2821.  The

   local parts of those addresses MAY be made up of any ASCII

   characters, although some of them MUST be quoted as specified there.

   It is notable in an internationalization context that there is a long

   history on some systems of using overstruck ASCII characters (a

Yao & Mao                     Experimental                     [Page 20]

RFC 5336                   EAI SMTP Extension             September 2008

   character, a backspace, and another character) within a quoted string

   to approximate non-ASCII characters.  This form of

   internationalization SHOULD be phased out as this extension becomes

   widely deployed, but backward-compatibility considerations require

   that it continue to be supported.

A.5.  Applicability of SMTP Extension to Additional Uses

   Among other protocol changes, the SMTP extension allows an optional

   alternate address to be supplied with the MAIL and RCPT commands.

   For the purposes of this set of specifications, this alternate

   address only has meaning when the primary address contains UTF-8

   characters and the message is downgraded.  While it may be tempting

   to consider the alternate address as a general-purpose second-chance

   address to be used whenever the primary address is rejected, such

   behavior is not defined here.  This restriction allows for future

   extensions to be developed which create such a general-purpose

   second-chance address, although no specific work on such an extension

   is currently anticipated.  Note that any such extension needs to

   consider the question of what the [RFC0974] sequencing rules mean

   when different possible servers support different sets of ESMTP

   options (or, in this case, addresses).  The answer to this question

   may also imply updates to [RFC2821].

Authors' Addresses

   Jiankang YAO (editor)

   CNNIC

   No.4 South 4th Street, Zhongguancun

   Beijing

   Phone: +86 10 58813007

   EMail: yaojk@cnnic.cn

   Wei MAO (editor)

   CNNIC

   No.4 South 4th Street, Zhongguancun

   Beijing

   Phone: +86 10 58812230

   EMail: maowei_ietf@cnnic.cn

Yao & Mao                     Experimental                     [Page 21]

RFC 5336                   EAI SMTP Extension             September 2008

Full Copyright Statement

   Copyright (C) The IETF Trust (2008).

   This document is subject to the rights, licenses and restrictions

   contained in BCP 78, and except as set forth therein, the authors

   retain all their rights.

   This document and the information contained herein are provided on an

   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS

   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND

   THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS

   OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF

   THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED

   WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Intellectual Property

   The IETF takes no position regarding the validity or scope of any

   Intellectual Property Rights or other rights that might be claimed to

   pertain to the implementation or use of the technology described in

   this document or the extent to which any license under such rights

   might or might not be available; nor does it represent that it has

   made any independent effort to identify any such rights.  Information

   on the procedures with respect to rights in RFC documents can be

   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any

   assurances of licenses to be made available, or the result of an

   attempt made to obtain a general license or permission for the use of

   such proprietary rights by implementers or users of this

   specification can be obtained from the IETF on-line IPR repository at

   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any

   copyrights, patents or patent applications, or other proprietary

   rights that may cover technology that may be required to implement

   this standard.  Please address the information to the IETF at

   ietf-ipr@ietf.org.